This project is an extension of previous work involving the creation, development, and testing of image processing techniques designed to improve diagnostic performance. Recent emphasis has centered on position-invariant filtering schemes to eliminate known sources of noise in CAT scans. Other work involves testing of the computer implemented transformation which creates an isometric three-dimensional "pseudo-solid" from any two-dimensional array of grey levels. Comparisons of processed and controlled images by unbiased observers suggest that diagnostic performance is limited by sample restrictions imposed by a severe angular dependence on the pseudo-solid algorithm when using shadows as cues to depth. Modifications of this and the other promising enhancement schemes reported on last year are being explored to further define factors limiting diagnostic performance. This project will continue to forgo investigation of empirical enhancement schemes per se in favor of approaches which are based on relatively specific theoretical concepts.